Draw in the WebGIS, and a free map while you edit

Until yesterday the WebGIS could only receive files: to correct the boundary of a parcel or add a missing shape you had to go back to your GIS software, export again and upload everything from scratch. The toolbar on the left of the map now holds a real drawing kit: point, line, polygon, rectangle and circle, plus editing the vertices of a shape that is already there and deleting the ones you no longer need. There is also automatic snapping to existing vertices, so two neighbouring areas match exactly instead of leaving gaps or overlaps of a few centimetres. Every shape you close is saved straight away on the layer selected in the list on the left, so there is no "save" button to remember at the end. If no layer happens to be selected at that moment, the platform creates one called "Sketches" and puts the drawing in it: you can rename and restyle it later, at your own pace. Even two shapes drawn one right after the other in quick succession end up in the same layer, with no duplicates. If the save does not go through — it happens on an unstable connection — the shape stays on the map and a message warns you that it was not saved, because throwing away your work over a network glitch would be the worst possible choice. Moving vertices works on a shape belonging to a layer other than the highlighted one too: the change is attributed to the right layer, the one that actually holds the shape. You will find the new tools under XNatura → Monitoring → Site → WebGIS, in the left-hand column above the list of layers.

When you correct the land use by hand, the drawing session used to keep the map completely still: if the site did not fit on screen, the only way to reach the corner you cared about was to close the editing session, move the map and start over. Now the hand — the tool that does not paint — hands dragging back to the map, while the wheel, the pinch on the touchpad and the + and − buttons work with any tool armed, brush and eraser included. Brush and eraser still lock dragging, and for good reason: while you are painting a press of the mouse must be able to mean one thing only, otherwise every attempt to move the map would leave an unwanted stroke behind. The delicate part was preserving the work already done: the canvas you paint on only covers the portion of map you can see, so on every move the strokes just laid down are folded into the full session image and the canvas is rebuilt from that. The result is that you can paint one corner, travel to the other side of the site, paint again and find everything in place, with no offset copies and no cover eating into its neighbours. Undo and redo cross the moves too: the history is a single sequence in which strokes and movements alternate in the order you worked, so you can step back several times even if the map has moved in the meantime. Before this change, a simple move cleared the history and left both buttons dead, freezing the work half done. While the map is in flight the canvas is hidden and reappears only once the new framing is ready, so you never see a slice of colour resting on ground it does not describe. Try it under XNatura → Monitoring → Site → Land use → Manual editor, by picking the hand in the toolbar.

The slider that sets how transparent the land-use layer is already existed on the map, but it stayed outside the editing session: as soon as you opened the manual editor there was no way left to look at the satellite imagery under the colours, and working out exactly where a row of trees ends or where the wood begins became an exercise in memory. The slider is now available inside the editor too, at the top left of the editing map, always within reach while you paint. It is the same control you use outside, not a second mechanism: the value you set on the map is the one you find on entering the editor, and the other way round, so there is nothing to set again every time you go in or out. Pulling it down makes the colours of the other covers progressively transparent and the satellite imagery emerges underneath, while the cover you are painting at that moment stays clearly visible: that is exactly the one you need in order to compare your stroke with the real ground. So you can trace a boundary using the aerial image as a guide, raise the opacity to check the overall result and lower it again to refine the details, without ever leaving the session. The top-left position is a deliberate choice, to stay clear of the zoom controls at the bottom: the two no longer steal each other's clicks. Transparency is only a way of displaying: the colour you save stays at full strength, so there is no risk of delivering a washed-out map because you had left the slider down. You will find the control under XNatura → Monitoring → Site → Land use → Manual editor, at the top left above the editing map.

The same report reached us more than once: "the opacity slider does not work, neither inside editing nor outside". There were two causes, different from each other but with the same discouraging effect. The slider inside the manual editor ignored the mouse entirely and answered the arrow keys only, because the measure that stops the map from panning during a drag ended up intercepting the gesture meant for the slider as well. Dragging is now tracked continuously, the value changes as you move the mouse and the map underneath stays exactly where you left it. The slider on the land-use map, on the other hand, did work but was a mere six pixels wide: the area to aim at was so thin that most attempts missed it by a hair, and the control looked inert. We gave it a column of twenty-four pixels to grab, without changing anything in its looks: the bar appears exactly as before, it simply gets caught on the first try. The third piece concerns what happens when you lower the transparency while editing: the cover you are working on no longer fades along with the others, so pulling the slider to zero leaves you with exactly the colour you are painting resting on bare satellite imagery, which is the whole point of this control. Before, both faded away and comparing with the ground became impossible. There is nothing to do to get it: open XNatura → Monitoring → Site → Land use and try dragging the slider, inside and outside the manual editor.

Renaming a land-use version could make it vanish from the list, and it was a particularly unpleasant kind of vanishing because it looked successful: the platform confirmed with a success message, the list appeared unchanged, and only on reloading the page did you notice that the version was gone. The reason is that version numbers are counted per name: giving a version a name already used by another family sent it into a numbering where a higher number already existed, so it ended up behind the others, outside the visible list. Renaming with an existing name now explicitly means "become the newest version of that name", and the numbering is updated accordingly, so the version stays where you expect to find it. To make all of this clear before any data is written, clicking "Rename" in the version's menu opens a "How it works" window with a before/after diagram: you read the name and number the version has now, which name it will take, and — when the chosen name is already in use — which entry will leave the list and where to find it again, namely in the CLC history. The explanation now arrives at the start, when you click the menu item, rather than at the end after pressing Save: by then you had already decided and reading it served little purpose. Then, when you pick a name that is already taken, a compact note inside the picker tells you precisely which version you are about to become and which one moves into the history. We also fixed the name search, which used to match the English term only: typing "prairie" in your own language found nothing, whereas the search now recognises the labels in every language of the platform. Finally we removed from the diagram the number that could not be predicted with certainty, because a wrong number stated confidently is worse than one number fewer. You will find all of this under XNatura → Monitoring → Site → Land use → CLC history, from a version's menu.

The window showing the visual comparison between land-use versions had a glaring layout flaw: the image and the information around it spilled out of the window's own edges. In practice the map ended up over the header, the caption and the opacity slider crossed the bottom edge, and the legend with the class names was pushed so far down that it was cut away entirely — precisely the part you need in order to know which colour stands for which cover. Every element now stays in its own band — the header at the top, the content in the middle with its own scrollbar, the controls at the bottom — and the legend is fully readable again. We also removed a fixed height that caused trouble on wide but short screens, such as thirteen- and fourteen-inch laptops: at those proportions the window overflowed symmetrically at the top and the bottom, and the tabs for switching view along with the close button ended up off screen, out of reach. The window's height now adapts to the browser's, so the controls stay visible and clickable whatever your resolution. Content that needs to scroll scrolls inside its own area, without dragging the rest of the window along. If in the past you gave up on this comparison because it was unusable on your screen, it is genuinely worth another go now. You open it from XNatura → Monitoring → Site → Land use → CLC history, by choosing the version comparison view.

In the WebGIS you could set colours, thicknesses and transparencies for your layers, but that styling work was only kept for the length of the visit: reloading the page, or coming back the next day, was enough to send everything back to the starting colours. For anyone who uses the map as a reading tool — one colour for the areas under conversion, another for the monitoring points, another for the boundaries — it meant redoing the same work every time. The style is now saved on the layer as you change it and read back on the next load, so the map greets you exactly as you left it. Saving is deliberately delayed by a fraction of a second while you drag the sliders, so moving a transparency back and forth does not generate dozens of pointless saves. In the same round we fixed four inaccuracies on the same page: the indicator at the bottom left always showed the same zoom level instead of following the map, an export that failed ended in silence without saying anything, a freshly drawn layer was sometimes mistaken for a satellite image and treated as one, and the two delete confirmations used the browser's grey little box instead of the platform's own style. Zoom and coordinates now genuinely follow the map, a failed export tells you so, the layer type is recognised correctly and the delete confirmations look like the rest of the interface. We also made the attribute table quicker, as it used to repeat the calculation of derived columns twice for every row. No action is needed on your side: open XNatura → Monitoring → Site → WebGIS, give your layers the colours you prefer and reload the page to see them stay in place.

The assistant that helps you write the report can modify the document's blocks at your request, and at the end of every operation it shows a summary with an outcome label. The trouble was that the summary counted the changes requested rather than the ones that went through: if a request could not be applied — because the block had been deleted in the meantime, say, or because that kind of reordering is not provided for — it was discarded in silence and the label still said "Applied". Anyone reading "1 block updated" while seeing nothing change in the document concluded, very reasonably, that the assistant does not carry out requests. Only what was genuinely applied is counted now: the summary lists the discarded operations with the reason why they did not go through, and the label becomes "Discarded" when nothing got in. Even the worst case, an entire series of changes rejected because it was malformed, is no longer mute: it used to end without a word, indistinguishable from an assistant sitting on its hands. In the same round we made the panel's secondary texts legible: the message time, the indicator signalling work in progress, the operation's name, the remaining credit and the hints were written in transparency on a dark background, with a contrast below half the recommended threshold for small text. The time, in particular, was practically invisible. Those same elements now use solid, lighter colours from the same palette, so they read effortlessly even on a dim screen or in a brightly lit room. You will find the chat under XNatura → Monitoring → Site → Report, by opening a document and using the assistant panel on the right.

The Risks and Opportunities page is extremely dense with indicators, and next to each one there is a "?" that is supposed to explain what you are looking at. Those explanations, however, often went round in circles ("impact level: the numerical value corresponding to the impact risk"), used technical terms without unpacking them — a percentile without saying a percentile of what — and changed style from one indicator to the next. We have rewritten them all, one by one, with the same plain-language approach: first what the indicator measures, with its unit, and then how to read the value you see, that is when it is high, when it is low and what it means for your site. The wall of text in the in-depth tooltip has become a five-step chain that follows the reasoning from beginning to end: from the environmental variables to the physical risk, then the sensitivity of the activity, the impact risk and finally the economic value exposed. On top of the rewrite we ran a second pass, this time a scientific check, comparing every text against the official indicator documentation: seventy-two explanations contained an error or an imprecision and have been corrected. Among the most significant corrections: some scales were described the wrong way round, so a value indicating well-preserved nature appeared to indicate its degradation and a more vulnerable coast looked safer; the probability of financial risks is not a percentage but a qualitative scale over four time horizons; the thresholds for extreme heat days refer to the mean daily temperature and not the maximum; wind risk measures the chronic stress linked to average speed, while exceptional gusts are a separate indicator. The Methodology page has been aligned too, because it described the matrix combining physical risk and sensitivity with a different number of levels from the one actually used in the calculation. All this work is already live in Italian, English, Spanish, French and German, so the explanation you read is the same whichever language you use. Hover over the "?" marks in XNatura → Monitoring → Risks and Opportunities to read the new explanations.

Creating a new cover type inside the manual land-use editor asked for four clicks and a hover before a category could even be chosen: add the type, ask for a row, discover the pencil hidden behind that blank row's hover state, and only then open the list of categories. Three of those steps asked nothing and said nothing, they were simply there to be done. Now creating a type brings its first row with it, that row opens straight onto the category picker, and the picker appears with the cursor in the search field, so you can narrow the list by typing the first letters the instant it shows up. "Cancel" behaves as you would expect too: on a row whose category you had not chosen yet it removes the row, and it takes the type away with it when the type had just been created, so "+" followed by "Cancel" leaves the list exactly as it was. Previously that row stayed there, blank and untitled, and blocked the possibility of defining a land use on that type for the rest of the session. We then fixed a behaviour that was disorienting: if you choose for one type a land use that another type already has, the two are the same thing and get merged, so the card you were filling in closes and the selection moves to the type that already had that land use. The point is that this card could sit anywhere in the list, even off screen: yours seemed to vanish into thin air and it looked as though nothing had happened. The list now always brings into view the card the selection has moved to, and it does not move when the card is already visible, so clicking a pencil never makes the list jump. Finally the category picker no longer loses the text you typed while the editor refreshes, and screen reader users hear the chosen category announced correctly. You will find all of this under XNatura → Monitoring → Site → Land use → Manual editor, by adding a type with the "+" button.